ck

Persist per-project session context and git activity for Claude Code.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/Maelwalser/claude-config --skill ck-maelwalser
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ck
Source: https://github.com/Maelwalser/claude-config/tree/main/skills/ck
Command: npx skills add https://github.com/Maelwalser/claude-config --skill ck-maelwalser

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Developers frequently lose conversational context, task state, and decisions across Claude Code sessions when switching projects or restarting the assistant. ck provides a persistent, per-project memory store that auto-loads context at session start, tracks session summaries and git activity, and ensures deterministic saves so progress is recoverable and searchable.

Core Features & Use Cases

  • Automatic session injection: a SessionStart hook injects a compact project briefing on every session start.
  • Deterministic CLI commands: register projects, save sessions, resume briefings, list projects, view quick info, forget contexts, and migrate v1 data using Node.js scripts.
  • Persistent storage & native memory: structured context.json + generated CONTEXT.md live under ~/.claude/ck/contexts and native memory entries are written for integration with other tools.
  • Use Case: a developer switching between multiple repositories can resume work instantly with the last session summary, left-off notes, next steps, and git-aware activity indicators.

Quick Start

Run /ck:init to register the current project and then run /ck:save after a session to persist the session summary and context.

Frequently Asked Questions about ck

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I persist Claude Code session context across project restarts?

Persistent per-project session context is saved by running /ck:save, which stores session summaries, left-off notes, and git-aware activity in structured context.json and CONTEXT.md files under your local ~/.claude directory for automatic reloading.

How does automatic session injection work for local Claude Code projects?

Automatic session injection uses a SessionStart hook to load a compact project briefing into Claude Code on every session start, utilizing deterministic Node.js scripts to read your saved context and restore your exact task state instantly.

Can I track git activity alongside session summaries for multiple repositories?

Git-aware activity indicators and session summaries are tracked together for multiple local developer repositories, managed via a projects.json registry that lets you list registered projects and resume work with the exact context you left off.

What is the best way to manage per-project memory for Node.js CLI projects?

Per-project memory management is handled by deterministic Node.js scripts executed from ~/.claude/skills/ck/commands, writing native memory entries and structured context files to the local filesystem for reliable progress recovery and searchability.

How do I migrate old context data when updating my session tracking setup?

Context data migration is supported via a dedicated CLI command that upgrades v1 data formats to the current structure, ensuring your historical session summaries and project configurations remain intact in the updated context.json store.

Does session tracking require any external dependencies to run locally?

No external dependencies are required for session tracking, as the Skill operates entirely through local filesystem and git access using built-in Node.js scripts, requiring only the Claude Code environment and local developer repositories.